PCCP

View Article Online

PAPER View Journal
Published on 03 October 2014. Downloaded by UNIVERSIDAD DE EXTREMADURA on 21/10/2014 10:40:58.

Electrical conductivity of activated carbon–metal

oxide nanocomposites under compression:
Cite this: DOI: 10.1039/c4cp03952a
a comparison study
A. Barroso-Bogeat,*a M. Alexandre-Franco,a C. Fernández-González,a
A. Macı́as-Garcı́ab and V. Gómez-Serranoa

Received 3rd September 2014,
Accepted 2nd October 2014
DOI: 10.1039/c4cp03952a
www.rsc.org/pccp
From a granular commercial activated carbon (AC) and six metal oxide (Al2O3, Fe2O3, SnO2, TiO2, WO3 and
ZnO) precursors, two series of AC–metal oxide nanocomposites were prepared by wet impregnation,
oven-drying at 120 1C, and subsequent heat treatment at 200 or 850 1C in an inert atmosphere. Here, the
electrical conductivity of the resulting products was studied under moderate compression. The influence of
the applied pressure, sample volume, mechanical work, and density of the hybrid materials was thoroughly
investigated. The DC electrical conductivity of the compressed samples was measured at room temperature
by the four-probe method. Compaction assays suggest that the mechanical properties of the nanocomposites
are largely determined by the carbon matrix. Both the decrease in volume and the increase in density were
relatively small and only significant at pressures lower than 100 kPa for AC and most nanocomposites.
In contrast, the bulk electrical conductivity of the hybrid materials was strongly influenced by the intrinsic
conductivity, mean crystallite size, content and chemical nature of the supported phases, which ultimately
depend on the metal oxide precursor and heat treatment temperature. The supported nanoparticles may
be considered to act as electrical switches either hindering or favouring the effective electron transport
between the AC cores of neighbouring composite particles in contact under compression. Conductivity
values as a rule were lower for the nanocomposites than for the raw AC, all of them falling in the range of
semiconductor materials. With the increase in heat treatment temperature, the trend is toward the
improvement of conductivity due to the increase in the crystallite size and, in some cases, to the formation of
metals in the elemental state and even metal carbides. The patterns of variation of the electrical conductivity
with pressure and mechanical work were slightly similar, thus suggesting the predominance of the pressure
effects rather than the volume ones.

1. Introduction
In recent years, novel composites and hybrid materials based
on metal and metal oxide nanoparticles (i.e., particles having
size in the range from 1 to 100 nm) supported on carbon
materials have received a great deal of attention owing to their
wide range of potential applications in heterogeneous catalysis,
photocatalysis, gas and liquid adsorption, solid state gas sensors,
electronics, energy storage devices, solar cells, and so on.
Among the great variety of carbon materials, activated carbon
(AC, henceforth) has been long selected as a support of metal-
containing nanoparticles because of its excellent and unique
a
Department of Organic and Inorganic Chemistry, Faculty of Sciences,
University of Extremadura, Avda. de Elvas s/n, E-06006 Badajoz, Spain.
E-mail: adrianbogeat@unex.es; Fax: +34 924 271449; Tel: +34 924 289421
b
Department of Mechanical, Energetic and Materials Engineering,
University of Extremadura, Avda. de Elvas s/n, E-06006 Badajoz, Spain
properties, which have been the subject of several reviews and
chapters in books.1–5 Supported nanoparticles include not only
metal oxides such as TiO2, Fe2O3, ZnO, WO3 and SnO2, but also
metal hydroxides and oxyhydroxides (Ni(OH)2,6,7 Zn(OH)2,8 and
a-FeO(OH)9,10), and metals in the elemental state (Pt,11 Au,12
Ag,13 Pd,14 Ni,15 Cu,16 and so on). A comprehensive review on
the preparation methods and applications of several metal
oxides supported on AC has been recently reported elsewhere.17
In addition to the intrinsic properties of the single constituents,
both the grain boundaries and the interfaces play a key role in
the overall chemical and physical properties of the aforesaid
carbon-based nanocomposites.18 Among other factors, electrical
conductivity is well known to largely determine the application
and performance of these materials as electrode materials in
energy storage devices, such as supercapacitors,19,20 lithium ion
batteries21,22 and fuel cells.23 On the other hand, several studies
have found a correlation between the electrical conductivity of a
wide range of metal oxides and their catalytic behaviour and/or

This journal is © the Owner Societies 2014 Phys. Chem. Chem. Phys.
 View Article Online

Paper PCCP

gas sensing properties.24–29 Therefore, it becomes apparent that discussed in terms of the content, chemical nature and crystal-
the measurement and comparison of the electrical conductivity lite size of the supported nanoparticles, and of the texture of
of carbon-based composites containing nanoparticles of different the nanocomposites, and on the basis of Holm’s contact theory.
metal oxides are key issues in order to assess many of their
potential applications.
Nevertheless, in the case of granular, powder and porous
2. Methodology
Published on 03 October 2014. Downloaded by UNIVERSIDAD DE EXTREMADURA on 21/10/2014 10:40:58.

materials such as AC–metal oxide nanocomposites, the deter- 2.1. Materials

mination of the electrical conductivity is rather complicated. A granular AC purchased from Mercks (Darmstadt, Germany),
According to the contact theory established by Mrozowski and of 1.5 mm average particle size (Cod. 1.02514.1000), as received
Holm30–32 more than sixty years ago, the electrical conductivity of a without further treatment, was used as a support of a broadly
granular or powder carbon material depends not only on the varied series of metal oxides, including Al2O3, Fe2O3, SnO2,
interparticle separation distance,33–38 but also on the mean particle TiO2, WO3 and ZnO. The selected metal oxide precursors were
size,39 as well as on the chemical nature of the carbon surface. Al(NO3)39H2O, Fe(NO3)39H2O, SnCl22H2O, Na2WO42H2O and
In this connection, the separation of the contribution of particle Zn(NO3)26H2O, all of them supplied by Panreacs (Barcelona,
cores and surfaces to the bulk conductivity of several powder Spain) and being of reagent grade. As an exception to the rule,
samples was methodically analyzed by Braun and Herger.40 anatase powder from Aldrichs (Steinheim, Germany), less than
Electron tunnelling, which entails the ability of electrons to 325 mesh average particle size, was used as a TiO2 precursor.
jump across the gaps and voids between closely spaced grains
and particles,41 may be considered as the predominant electrical 2.2. Preparation of AC–metal oxide nanocomposites
transport mechanism in granular and powder carbon materials. The preparation of the AC–metal oxide nanocomposites was carried
Consequently, the narrowing of gaps and the increase in contact out following the methods previously described in detail by Barroso-
between particles should lead to a significant increase in the Bogeat et al.,52,53 which are briefly summarized in Table 1 together
bulk electrical conductivity of the sample. Since the pioneering with the codes assigned to the resulting products. Accordingly, two
studies conducted by Skaupy and Kantrowicz42 on the electrical series of hybrid materials were prepared depending on heating
conductivity of compressed metal powders, many studies conditions, i.e. series 1 (200 1C, 2 h) and series 2 (850 1C, 2 h).
concerning the measurement of the conductivity of granular
and powder materials under compression, in particular 2.3. Characterization of AC–metal oxide nanocomposites
carbons,30,36,43–47 have been reported. 2.3.1. Ash content. The ash content of AC and hybrid
The number of contacts established between neighbouring materials was determined by incineration at 650 1C for about
particles is of major relevance, as it determines the number of 12 h in air, using a muffle furnace (Selectas). The soaking time
channels or paths available for the effective transport of electric at maximum heat treatment temperature was considered to be
current.41 Several theoretical studies considering the particles long enough to achieve the complete oxidation and conversion
as equal or roughly equal spheres in a random packing48–51 of the material fraction contained in each sample into gaseous
have been performed in order to estimate the degree of contact products and residual ashes. The results were recently reported
between them during compaction. However, real situations are elsewhere,53 and are collected in Table 2.
considerably more complicated, the mean number of contacts 2.3.2. X-ray diffraction (XRD). As described before,53 the
being almost impossible to determine since the particles may crystalline phases present in each nanocomposite were inves-
exhibit simultaneously a wide distribution of sizes and a strongly tigated by powder XRD, using a D8 Brukers Advance diffractometer
variable morphology,44 which even may undergo changes during
the compacting step.
In the present work, the electrical conductivity is studied Table 1 Methods of preparation and sample codes for AC–metal oxide
under moderate compression (i.e., from 99.5 to 756.0 kPa) for nanocompositesa
two series of AC–metal oxide nanocomposites, which were
Series Precursor MHTT/1C t/h A Code
previously prepared by our research group from a commercial
granular AC and six metal oxide (Al2O3, Fe2O3, SnO2, TiO2, WO3 Series 1 Al(NO3)39H2O 200 2 N2 A200
and ZnO) precursors by wet impregnation, oven-drying at 120 1C Fe(NO3)39H2O 200 2 N2 F200
SnCl22H2O 200 2 N2 S200
and subsequent heat treatment at two very different temperatures, TiO2 (anatase) 200 2 N2 T200
200 and 850 1C, in an inert atmosphere.52,53 These temperatures Na2WO42H2O 200 2 N2 W200
Zn(NO3)26H2O 200 2 N2 Z200
were selected in order to investigate and compare the influence of
the heating conditions on the electrical conductivity of the Series 2 Al(NO3)39H2O 850 2 N2 A850
nanocomposites as a result of the phase composition changes Fe(NO3)39H2O 850 2 N2 F850
undergone for each metal oxide precursor during the heat treat- SnCl22H2O 850 2 N2 S850
TiO2 (anatase) 850 2 N2 T850
ment stage. The influence of the applied pressure, volume Na2WO42H2O 850 2 N2 W850
change under compression, mechanical work, and density on Zn(NO3)26H2O 850 2 N2 Z850
the electrical conductivity is investigated for the resulting carbon- a
Abbreviations: MHTT, maximum heat treatment temperature; t, iso-
based hybrid materials. The obtained results are thoroughly thermal time at MHTT; A, atmosphere.

Phys. Chem. Chem. Phys. This journal is © the Owner Societies 2014
 View Article Online

PCCP Paper

Table 2 Ash content, crystalline phases, and their weight fraction (W), mean crystallite size (D), density (r) and bulk electrical conductivity at room
temperature (sRT) for the prepared AC–metal oxide nanocomposites

Sample Ash content/wt% Crystalline phase W/% D/nm r/g cm3 sRT/S m1
A200 8.86 Alumina (Al2O3) 9.9 1.88 4.0–4.1 1012 (ref. 58)

F200 13.55 Hematite (a-Fe2O3) 23.4 3.22 5.3 4.76  104 (ref. 59)
Published on 03 October 2014. Downloaded by UNIVERSIDAD DE EXTREMADURA on 21/10/2014 10:40:58.

Goethite (a-FeO(OH)) 26.3 3.17 3.8 5  105 (ref. 60)

Magnetite (Fe3O4) 9.6 6.19 5.2 2.25  104 (ref. 61)

S200 27.21 Cassiterite (SnO2) 16.0 4.05 6.9 1.9  104 (ref. 62)
Romarchite (SnO) 5.0 5.07 5.5 5.40  104 (ref. 63)

T200 8.05 Anatase (TiO2) 11.4 3.53 3.9 6.71  103 (ref. 64)

W200 9.93 Tungsten oxide (WO3) 4.4 2.37 7.2 1.3  106–7.5  106 (ref. 65 and 66)
Tungstite (WO3H2O) 1.4 3.05 5.5 6.7  102 (ref. 67)

Z200 9.45 Zinc oxide (ZnO) 0.9 20.80 5.6 4.8  108–1.3  106 (ref. 68)
Zinc hydroxide (Zn(OH)2) 7.3 1.89 3.1 o109 (ref. 8)

A850 5.68 Corundum (a-Al2O3) 11.0 2.06 4.0

F850 14.66 Hematite (a-Fe2O3) 41.2 36.60 5.3 4.76  104 (ref. 59)
Maghemite (g-Fe2O3) 5.9 15.11 4.9 2.85  105 (ref. 69)
Iron (Fe) 22.0 7.9 1.0  107 (ref. 70)

S850 2.34 Cassiterite (SnO2) 3.3 36.16 6.9 1.9  104 (ref. 62)
Tin (Sn) 36.1 89.62 7.4a/5.8b 9.9  106 (ref. 70)

T850 9.33 Anatase (TiO2) 7.3 2.31 3.9 6.71  103 (ref. 64)
Rutile (TiO2) 7.1 1.86 4.3 o107 (ref. 71)

W850 11.54 Tungsten (W) 9.0 48.36 19.3 1.9  107 (ref. 70)
Tungsten carbide (W6C2.54) 7.8 3.67 5.0  104 (ref. 72)

Z850 6.18 Zinc oxide (ZnO) 0.6 69.38 5.6 4.8  108–1.3  106 (ref. 68)
a b
White tin. Grey tin.

working with Cu Ka radiation (l = 1.5406 Å). The identification
of the crystalline phases was accomplished by comparing peak
positions and intensities with standard JCPDS files and with
data previously reported in the literature. The mean crystallite
size (D) was estimated for each of the crystalline phases in the
nanocomposites by applying the Scherrer’s equation54,55 to
their corresponding diffraction peaks:
K l
D¼
b  cos y
where K is the so-called Scherrer’s constant (K = 0.9454), l is the
wavelength of X-ray radiation, b denotes the full-width at half-
maximum of the diffraction peak and y is the characteristic
Bragg angle of the crystalline phase.
The powder XRD patterns recorded for both series of nano-
composites are shown in Fig. 1. The crystalline phases sup-
ported on AC for each composite, as well as their corresponding
weight fraction and mean crystallite size, have been reported in
a previous work,53 and are compiled in Table 2. The electrical
conductivity of the phases covering the AC particles is very
difficult to know, as it strongly depends on a number of factors
such as the amorphous or polycrystalline nature of the phase,
the presence of vacancies or dopant impurities, and so on.56,57
Fortunately, the effects of the aforementioned factors usually
alter only the conductivity values of the bulk phases by less
than one order of magnitude, which can be found tabulated in
specialised literature and have also been included in Table 2.
2.3.3. Textural data. The overall resistivity of granular and
powder composites may be regarded as the sum of all the
intraparticle and interparticle resistances. The intraparticle
resistances are in turn determined by the intrinsic resistivity
of the two single constituents (i.e., supported nanoparticles and
AC) and the resistance of the interface existing between them.73
(1) The resistance of each composite is largely a function of the
matter-free space in the sample. This space comprises not only
the interparticle voids, but also the porosity corresponding to
the intraparticle voids. It is well known that in porous solids,
such as activated carbons and carbon blacks, the microporosity
(i.e., the porosity region made up of pores of width smaller
than B 2 nm74) usually has the greatest contribution to the
surface area. Herein, the specific surface area (SBET) of AC and
the prepared nanocomposites was estimated by applying the
Brunauer, Emmet and Teller (BET) equation75 to the N2 adsorption
isotherms measured at 196 1C in a semiautomatic adsorption
equipment (Autosorb 1, Quantachromes). The theoretical back-
ground for microporosity characterization was based on Dubinin’s
theory. Thus, the application of the Dubinin-Radushkevich
(D-R) equation to the adsorption isotherms led to the values of
the micropore volume (W0).76 The mesopore (Vme) and macro-
pore (Vma) volumes were derived from the mercury intrusion

This journal is © the Owner Societies 2014 Phys. Chem. Chem. Phys.

Paper
Published on 03 October 2014. Downloaded by UNIVERSIDAD DE EXTREMADURA on 21/10/2014 10:40:58.
Fig. 1 Powder XRD patterns of nanocomposites of series 1 and 2.
curves obtained on a mercury porosimeter (PoreMaster-60,
Quantachromes). Finally, the total pore volume (VT 0 ) was
calculated from W0, Vme and Vma. These textural data for the
raw AC and the prepared hybrid materials have been previously
reported elsewhere,52 and are listed in Table 3.
2.4. DC electrical conductivity measurements
Based on a standard procedure previously reported,41,43,44,77 the
DC electrical conductivity (s) was measured at room temperature
Phys. Chem. Chem. Phys.
View Article Online
PCCP
by the technique of impedance spectroscopy over the frequency
range from 20 to 106 Hz at a voltage of 1 V, using the experi-
mental set-up schematized in Fig. 2.41 The masses of AC and
nanocomposites used, after oven-drying at 110 1C overnight,
were around 0.09 g, as seen in Table 3. Each accurately weighed
sample was then poured into a hollow thick-walled PVC cylinder
with an inner diameter of 0.8 cm, and compressed in air between
two close-fitting brass plungers forming the electrodes, the lower
one being fixed and the upper one movable. The compression
This journal is © the Owner Societies 2014
 View Article Online

PCCP Paper

Table 3 Textural data and sample mass for AC and prepared nanocomposites where R is the electrical resistance in O, A the area of the
0 plunger surface in m2, and h denotes the distance between the
SBET/ W0/ Vme/ Vma/ VT /
Series Sample m2 g1 cm3 g1 cm3 g1 cm3 g1 cm3 g1 m/g two plungers in the compression chamber, or the height of
the sample bed, in m. The resistance of the plungers was verified
AC 711 0.36 0.15 0.21 0.72 0.0924
and found to be at around 0.050 O, much lower than those of the
Series 1 A200 556 0.27 0.13 0.16 0.56 0.0908 samples under study. The electrical conductivity (s), volume (V)
Published on 03 October 2014. Downloaded by UNIVERSIDAD DE EXTREMADURA on 21/10/2014 10:40:58.

F200 618 0.29 0.13 0.14 0.56 0.0929 and density (r) were determined for each sample and at each
S200 302 0.15 0.08 0.13 0.36 0.0916
T200 776 0.35 0.13 0.19 0.67 0.0922 applied pressure (P). V was roughly estimated by considering
W200 646 0.32 0.15 0.17 0.64 0.0963 the compressed sample as a cylinder having a base of surface
Z200 651 0.31 0.18 0.20 0.69 0.0941
area A and height h:
Series 2 A850 759 0.34 0.14 0.19 0.67 0.0948
V = Ah (3)
F850 694 0.31 0.17 0.18 0.66 0.0974
S850 680 0.33 0.16 0.18 0.67 0.0910
T850 847 0.37 0.14 0.20 0.71 0.0912
while r was given by:
W850 741 0.35 0.15 0.20 0.71 0.0903 m m
Z850 834 0.32 0.15 0.18 0.65 0.0935 r¼ ¼ (4)
V Ah

3. Results and discussion

Fig. 2 Experimental assembly for measuring the electrical conductivity of
granular and powdered materials under moderate compression.
pressures ranged from 99.5 to 756.0 kPa, which were achieved
by applying different loads between 5 and 38 N on the upper
piston by a universal testing machine (Shimadzus), with a
precision of 1 mm in the compression plate movement. In this
regard, it should be noted that this range of applied pressures
was not very wide; the lowest pressure was enough to get a good
electrical contact between the sample particles and the plungers,
whereas the highest one was too low to cause the crushing and
breaking of the grains. Consequently, the increase in the volume
fraction of particles due to their spatial rearrangement is
expected to be the only effect of the compaction. The very small
changes produced in the height of the sample bed as a con-
sequence of the compression were obtained from the corre-
sponding stress–strain curve. The DC electrical resistance was
measured by the four-probe method43,45 using a digital multi-
meter (model 34411A, Agilents). After a series of preliminary
tests, 5 min were allowed prior to measuring the resistance in
order to avoid any transient effect of sample relaxation. The four
probes, i.e. IL, IH, VL, and VH, were connected to the electrodes as
illustrated in Fig. 2. Conduction was considered to be ohmic in
nature, so the electrical conductivity was estimated using the
following expression:30
h hybrid materials are likely to play a more important role
s¼
RA
3.1. Change in the sample volume under compression
Fig. 3(a) and (b) show the variation of the sample volume (V)
with applied pressure (P) for nanocomposites of series 1 and 2,
respectively. The plot for the raw AC has also been included in
the above figures for comparison purposes. V, as determined,
must be regarded as a very rough estimation of the bulk volume,
since the granular morphology of the samples leads to a loose
packing of the grains with voids between them, and not even at
the highest P value it can be ensured that the total void space has
been entirely excluded. Consequently, this volume is higher than
the bulk volume (i.e., that measured by mercury porosimetry78,79),
and decreases with increasing pressure to approach the latter.
It is seen first that for most nanocomposites the total reduction
in V with increasing P is relatively small, less than 7%, and
lower than that observed for AC (10.4%), being only significant
for samples Z200 and Z850 (11.0 and 9.7%, respectively). As a
rule, the decrease in V is more important at low P values, below
E100 kPa, while at higher pressures the variation is nearly
linear and negligible. Moreover, V is rather similar for AC and
most composites over the entire P range, and it varies in the
order: S200 { A200 o F200 E T200 E Z200 o AC o W200 for
samples of series 1, and by: A850 o S850 o W850 o Z850 E
T850 E AC { F850 for those of series 2. As seen in Table 3, the
hybrid materials prepared at 200 1C possess a very different
degree of development of their surface area, with SBET values
widely ranging from 302 m2 g1 for S200 to 776 m2 g1 for T200,
whereas the specific surface area is more homogeneous for
nanocomposites prepared at 850 1C, as it varies in the narrow
range between 680 m2 g1 for S850 and 847 m2 g1 for T850.
Therefore, the extent of surface area of the composites should
not be identified as the main factor in the magnitude of change
in V produced by compression. Neither is the microporosity
of the samples, since it is well known that for porous solids
the surface area is largely concentrated in the micropores.41
Conversely, the mesoporosity and the macroporosity of the
(2)
than the above factors, especially for composites of series 1.

This journal is © the Owner Societies 2014 Phys. Chem. Chem. Phys.

Paper
Published on 03 October 2014. Downloaded by UNIVERSIDAD DE EXTREMADURA on 21/10/2014 10:40:58.
Fig. 3 Sample volume versus applied pressure for nanocomposites of (a) series 1 and (b) series 2.
Indeed, from Table 3 one observes that Vme, Vma and VT 0 are
markedly lower for S200 than for AC and the rest of the samples
prepared at the same temperature, and the latter is higher as
follows: S200 { A200 E F200 o W200 o T200 o Z200 o AC,
which is in good agreement with the sequence obtained for the
variation of V with P. Nevertheless, such a correlation cannot
be established for composites of series 2, as VT 0 increases in
the following order: Z850 o F850 o A850 E S850 o T850 E
W850 o AC. As a rule, the slight differences in V with increas-
ing P for AC and most of the nanocomposites, irrespective of
their heat treatment temperature, may be explained by taking
into account that the VT 0 values fall within a very narrow range
from 0.56 cm3 g1 for A200 to 0.72 cm3 g1 for the carbon
support. The great contribution of VT 0 to V comes from the fact
that the former cannot be completely excluded from the latter
without crushing and breaking the grains, which usually
requires compression pressures much higher than those
applied in the present work. In this regard, it is worth noting
that for AC and most composites VT 0 accounts for more than a
third of V at any P value. Thus, for example, VT 0 is about 39% of
Fig. 4 Density versus applied pressure for nanocomposites of (a) series 1 and (b) series 2.
Phys. Chem. Chem. Phys.
View Article Online
PCCP
V for A850 without compression, a value which rises up to 41%
at P = 756.0 kPa.
3.2. Variation of q with P
The density is an intrinsic property of each hybrid material. The
variation of r with P for nanocomposites of series 1 and 2 is
depicted in Fig. 4(a) and (b), respectively, together with the
curve for AC. Similarly to the sample volume, the density, as
estimated in the present work, must be considered as a very
rough estimate of the bulk density. Such density is lower than
the apparent density, i.e. that measured in a typical mercury
porosimetry experiment,78,79 and increases with the applied
pressure to approach the latter. As can be seen from the above
figures, both AC and all the composites show a quite similar
densification-pressure behaviour, namely a linear dependence
of density on the applied pressure at P 4 100 kPa. Analogous
densification behaviour has been recently reported by Marinho
et al.46 for a variety of carbon materials, including graphene,
multiwalled carbon nanotubes, carbon black and graphite
powder, which they interpreted as follows. After pouring the
This journal is © the Owner Societies 2014
 View Article Online

PCCP Paper

sample into the hollow plastic cylinder, a loose packing with Table 4 Electrical conductivity of AC and prepared nanocomposites at
large voids between the particles and grains is obtained. Under two pressure values
moderate compression (i.e., within the range of pressures Series Sample s1/S m1 P = 100 kPa s2/S m1 P = 756 kPa PIECa
applied in this work), densification is mainly attributable to
AC 57.3 194.1 239
the rearrangement of grains controlled by the interparticle
friction and, thereby, no significant changes in the internal Series 1 A200 25.1 103.7 314
Published on 03 October 2014. Downloaded by UNIVERSIDAD DE EXTREMADURA on 21/10/2014 10:40:58.

porous structure of the single particles are expected. Moreover, F200 7.8 40.0 415
S200 23.3 162.3 597
for most nanocomposites the overall increase in r with increas- T200 52.2 172.0 229
ing P is very low, and more relevant at P o 100 kPa, in line with W200 38.5 161.1 319
that stated above for the variation of V. At higher P values, the Z200 35.8 149.1 317
variation is linear and the corresponding fitted lines have an Series 2 A850 40.9 150.8 269
almost identical slope for AC and all the hybrid materials. Over F850 36.9 131.5 256
the entire P range, the order of the variation of r is AC o W200 o S850 41.9 173.4 314
T850 44.4 172.5 288
T200 o F200 E Z200 o A200 o S200 for samples heated at W850 63.8 209.5 229
200 1C, and F850 o AC E T850 o W850 E Z850 o S850 o A850 Z850 48.3 157.2 226
for those prepared at 850 1C. If one assumes that the density of a
Percentage of increase in electrical conductivity.
the supported nanoparticles (see Table 2; here we have consid-
ered such density to be roughly equal to the bulk density,
although in the case of nanoparticles the density has been WO3 and ZnO), which are well known to be wide band gap
recently reported to be dependent on the particle size80) is as a semiconductors.17 Therefore, the presence of these oxides must
rule much higher than that of the raw AC (2.0 g cm3, measured decisively contribute to the semiconductor properties exhibited
by helium pycnometry52), then the density of the nanocomposites by the hybrid materials. In addition, the values of s1 and s2 are
should increase with their ash content, since the greater the as a rule markedly lower for the nanocomposites than for the
latter, the larger the amount of the aforesaid nanoparticles pristine AC, except for the sample W850. The greatest electrical
supported on AC. As seen in Table 2, the ash content increases in conductivity for this composite may be related to the presence of
the following order: T200 o A200 o Z200 o W200 o F200 o nanoparticles of elemental tungsten and the non-stoichiometric
S200 for samples of series 1, and by: S850 o A850 o Z850 o tungsten carbide W6C2.54 as unique supported metal phases, whose
T850 o W850 o F850 for those of series 2. Strikingly, a intrinsic conductivity values at room temperature (see Table 2) are
correlation of density with the ash content is not found; indeed, noticeably higher than for other metal phases. Table 4 also lists the
for composites heated at 850 1C the density behaviour is nearly percentage of increase in electrical conductivity (PIEC), which is
the opposite to the expected one according to the ash content defined as follows:
values. The differences between both variation sequences for ðs2  s1 Þ
density and ash content are likely connected with the density of PIEC ¼  100 (5)
s1
the supported inorganic phases, which is a very important
factor as it widely ranges between 3.1 g cm3 for Zn(OH)2 and 3.3.1. Variation of r with P. The plots of s against P for the
19.3 g cm3 for metallic W. hybrid materials prepared by heating at 200 and 850 1C are shown
Finally, it is worth mentioning that the similar behaviour of in Fig. 5(a) and (b), respectively. As expected, s increases with P for
V and r under compaction for AC and all nanocomposites, all samples, and a good linear relationship is also observed
regardless of their metal oxide precursor and heat treatment between them, irrespective of the heat treatment temperature. As
temperature, clearly suggests that the mechanical properties of stated above, the electrical conductivity strongly depends on the
the latter are mainly determined by the carbon matrix. number of effective electrical contacts which are established
between the sample particles, since this determines the number
3.3. Study of electrical conductivity of channels or paths available for the transport of electric current.
The values of s measured for the AC support and the pre- Under compression, the neighbouring particles are forced to
pared hybrid materials at two P values, P = 99.5 kPa (s1) and approach each other, readily increasing the number of electrical
P = 756.0 kPa (s2), are collected in Table 4. One notes that all contacts and hence the conductivity.30,41 A few previous studies
the samples show values of s1 and s2 falling in the range of have revealed that the electrical conduction under compression in
conductivity of typical semiconductor materials, i.e. from 108 different granular and powder carbon materials,30,45,47 especially
to 105 S m1.81 This fact should not be surprising, since several carbon blacks,41,90 is usually favoured by a smaller sample volume.
previous studies have pointed out the semiconductor behaviour Nevertheless, such an effect of sample volume on electrical con-
for a number of carbon materials, such as activated carbons ductivity is not observed in the case of these nanocomposites,
prepared from different lignocellulosic precursors by both chemical since over a wide P range s varies in the order: AC 4 T200 4
and physical activation methods,82–85 amorphous carbons86–89 S200 E W200 4 Z200 4 A200 4 F200 for 200 1C samples, and by:
and carbon-based composites.18 Furthermore, the nanoparticles W850 4 AC 4 T850 E S850 4 Z850 4 A850 4 F850 for 850 1C
detected by powder XRD in the prepared carbon-based hybrid samples. Neither a correlation between the above sequences and
materials are mostly metal oxides (Al2O3, Fe2O3, SnO, SnO2, TiO2, the total variation in V under compression is found.

This journal is © the Owner Societies 2014 Phys. Chem. Chem. Phys.

Paper
Published on 03 October 2014. Downloaded by UNIVERSIDAD DE EXTREMADURA on 21/10/2014 10:40:58.
Fig. 5 Electrical conductivity versus applied pressure for nanocomposites of (a) series 1 and (b) series 2.
As shown in Table 3 and previously discussed in detail
elsewhere,52 the preparation of AC–metal oxide nanocomposites
by wet impregnation of AC with the metal oxide precursor, oven-
drying at 120 1C, and subsequent heat treatment at 200 1C in an
inert atmosphere, leads to a decrease in the three porosity regions,
i.e. micro-, meso- and macroporosity, for the resulting products as
compared to the raw AC. Conversely, by heat treatment at 850 1C
porosity develops for all hybrid materials, the increase being more
pronounced in the region of micropores. However, the micropore
volume (W0) as a rule is significantly lower for these composites
than for the carbon support. In brief, both heat treatments at
200 and 850 1C cause a reduction of porosity in the three
regions of pore sizes, which is less marked for the nano-
composites prepared at the highest temperature. Unlike the
behaviour observed for other porous carbon materials, such a
reduction in porosity, and hence in the total intraparticle void
space (i.e., VT 0 ), does not result in an improvement of the
electrical conductivity of AC, since most hybrid materials exhibit
s values lower than the carbon support over the entire P range
(see Fig. 5(a) and (b)). As the only exception to this rule, the
conductivity is noticeably higher for W850. Therefore, it becomes
apparent that the porosity of AC is partially filled and blocked by
a variety of nanoparticles having a much lower conductivity than
the carbon support itself. As far as the composites of series 1
are concerned, such nanoparticles are essentially metal oxides,
whose electrical conductivity values at room temperature (see
Table 2) are generally several orders of magnitude lower than
that measured for the raw AC (sRT = 19.9 S m1). In this regard,
the conductivity of these hybrid materials should decrease with
increasing the content of nanoparticles and, thereby, with the
ash content (for its variation, see Section 3.2). However, this
correlation is not observed and, similarly to that stated above
for the variation of r with P, the reason may be attributable to
the very different intrinsic conductivity values of the supported
nanoparticles. Thus, the relatively low s values registered for
A200 are explained by the insulating character of the supported
alumina,58,91 even though such a composite should theoreti-
cally exhibit higher conductivity due to its low ash content.
Phys. Chem. Chem. Phys.
View Article Online
PCCP
Likewise, the highest ash content for S200 should give the poorest
s values for this hybrid material. Nevertheless, from Fig. 5(a) it is
seen that the s–P plot for S200 is situated significantly above than
those of F200 and A200. This fact seems to be related to the
much greater electrical conductivity of cassiterite (SnO2) and
romarchite (SnO) as compared to hematite (a-Fe2O3) and goethite
(a-FeO(OH)), and with alumina (Al2O3), which are the predominant
phases detected in S200, F200 and A200 (see Table 2).
From Fig. 5 and data in Table 4 it is evident that the increase
of heat treatment temperature from 200 to 850 1C as a rule
leads to an improvement of conductivity for the resulting
nanocomposites. This behaviour may be explained on the basis of
composition and structural changes in the carbon support and the
supported nanoparticles originated by the heat treatment at 850 1C
and with influence on electrical conductivity of the hybrid materials,
as seen below. First, as previously discussed elsewhere,52,53
carbothermic reductions of metal oxides occur during heating
at a temperature as high as 850 1C. As shown in Table 2, these
reactions give rise to the formation of nanoparticles of
elemental metals, mainly Fe, Sn and W, and even of metal
carbides, such as W6C2.54. In this connection, it should be
noted that a great variety of iron species, including several iron
oxides (FeO,92 g-Fe2O3,93 and Fe3O494) and carbides (Fe3C92 and
Fe4C92), as well as elemental iron,93 have been detected in
carbon–iron composites prepared under similar heating condi-
tions. However, only a-Fe2O3, elemental iron and a small amount
of g-Fe2O3 are detected in sample F850, these results being in
good agreement with those obtained in previous studies of
preparation of AC–iron oxide composites.95–98 The intrinsic
conductivity of the aforesaid metals and metal carbides is many
orders of magnitude higher than for the corresponding metal
oxides and, obviously, their presence in the nanocomposites
causes a significant enhancement in electrical conductivity.
Second, regardless of the metal oxide precursor, the heat
treatment at 850 1C usually results in the development of
crystallinity and in an increase of the crystallite size for the
supported nanoparticles,53 which also contribute to the better
conductivity exhibited by these composites. In this regard,
This journal is © the Owner Societies 2014
 View Article Online

PCCP Paper

a number of previous studies dealing with the dependence of
electrical properties on crystallite size for nanoparticles of a
wide variety of metal oxides99–103 and metals104 have shown
that the larger the crystallite size, the smaller the width of
the insulating grain boundary and, thereby, the greater the
electrical conductivity. Furthermore, from Table 4 one can infer
Published on 03 October 2014. Downloaded by UNIVERSIDAD DE EXTREMADURA on 21/10/2014 10:40:58.
for series 1, and S850 4 T850 4 A850 4 F850 4 AC 4 W850 E
Z850 in the case of series 2. The plots of s divided by r against
P for AC and the prepared nanocomposites are fairly similar to
those of s versus P in Fig. 5(a) and (b). Therefore, they have been
omitted for the sake of brevity.
On the other hand, based on an increase in the contact area

that the presence of zero-valent metal nanoparticles appears to between the particles and grains as a result of compression,
be a stronger factor than the increase of the crystallite size on Holm proposed the following exponential dependence of s on
the improvement of conductivity of the hybrid materials with P for powder and particulate materials:31,32
increasing heat treatment temperature. Indeed, the increase in
s = aPb (6)
conductivity is markedly greater for sample F850 and also,
though less, for S850 and W850, which are those containing where both a and b are constants whose values depend only on
metals in the elemental state, whereas such an increase is less the material properties. A value of b of about 0.5 has been
pronounced for the rest of the nanocomposites. As an exception associated with the elastic contact between the particles, whereas
to the rule, conductivity only decreases slightly for T850, likely for the plastic contact b has been estimated to be around 0.33.
due to the smaller crystallite size. It is worth noting that rutile The above equation was used to fit the experimental data shown
originates from anatase by means of a phase transition occur- in Fig. 5(a) and (b), and the results are gathered in Table 5.
ring in the temperature range from 600 to 700 1C for TiO2– It is seen that irrespective of the metal oxide precursor and the
carbon hybrid catalysts.105 As previously reported by Czanderna heat treatment temperature, the data as a rule fit well to Holm’s
et al.,106 this phase transition occurs extremely fast at tempera- equation, R2 values being above 0.99. The b values for AC and
tures above 730 1C. Therefore, a complete transformation of the all composites are much closer to 0.5 than to 0.33, especially for
initial pure anatase into rutile would be expected in the case of samples prepared at 850 1C, thus suggesting a nearly elastic
T850 due to the heat treatment at 850 1C for 2 h. Nevertheless, contact between the particles rather than plastic deformation.
the powder XRD pattern of T850 clearly reveals that only half The latter assertion is also corroborated from the corresponding
of the anatase is converted into rutile after such heat treatment. strain–stress curves, which exhibit the linear relationship typical
These results are in good agreement with those previously of the elastic behaviour over the entire range of applied pressures
obtained in the preparation of AC–TiO2 composites by the (these plots have been omitted for the sake of brevity). Moreover,
sol–gel method.107,108 The effect of the AC matrix on the TiO2 for hybrid materials of series 1, it also becomes apparent that the
phase transformation has been thoroughly investigated by deviation from the elastic deformation behaviour is greater for
Y. Li et al.108 They found that the temperature of the anatase those samples with higher ash content and, presumably, with a
to rutile phase transition in AC–TiO2 composites is higher than larger amount of supported nanoparticles. In fact, the highest
that observed in pure anatase powders. This behaviour was b value is obtained for S200, having an ash content of 27.21%.
attributed to the high surface area AC matrix, which alters the Nevertheless, a similar correlation has not been observed in the
phase transition due to its large interfacial energy, thus result- case of the composites prepared at 850 1C.
ing in anti-calcination effects for the AC matrix. In addition, the 3.3.2. Variation of r with volume. Similarly to density, the
small amounts of SiO2 in the AC support (around 1.2 wt%), and sample volume under compression may also be regarded as
coming from the inorganic matter present in the AC precursor, another intrinsic property of each hybrid material. Fig. 6(a)
have also been suggested to constrain the anatase to rutile and (b) reveal that similar s values are achieved with extremely
transformation, probably as a result of the formation of Ti–O–Si different sample volumes and, in the same way, identical sample
bonds during the heat treatment, which are much more diffi- volumes also lead to very different s values. For example, in the
cult to recombine with other Ti–O bonds, thus retarding the case of series 1, a conductivity of 80 S m1 is obtained with a
phase transition.109 sample volume of E0.123 cm3 for S200, 0.158 cm3 for A200 and
The conductivity s is markedly influenced by the applied 0.182 cm3 for W200, being E0.177 cm3 for AC. Concerning
pressure (see Fig. 5(a) and (b) again), especially for composites series 2, the above conductivity value requires a sample volume
of series 1. At low P values, up to E100 kPa, s is quite similar in of E0.158 cm3 for A850, 0.175 cm3 for Z850 and 0.192 cm3 for
value for all the composites prepared at 850 1C and slightly F850. Therefore, it is evident that the range of sample volumes
lower than for AC, except for W850, while the differences in
s are more significant in the case of the 200 1C samples.
At higher P, the greatest increase in s occurs for T200, W200 Table 5 Fit parameters for the Holm equation
and S200, and for W850, T850 and S850. The PIEC values Sample a b R2 Sample a b R2
collected in Table 4 clearly show that the effect of increasing
AC 3.32 0.61 0.9957 AC 3.32 0.61 0.9957
P on s is very strong for AC and both series of hybrid materials, A200 0.88 0.71 0.9900 A850 2.36 0.63 0.9967
being somewhat weaker for those prepared at higher temperature. F200 0.19 0.81 0.9997 F850 2.08 0.62 0.9985
Such values range between 229 for T200 and 597 for S200, and S200 0.32 0.95 0.9952 S850 1.45 0.72 0.9910
T200 3.56 0.58 0.9969 T850 2.18 0.66 0.9970
from 226 for Z850 to 314 for S850. The variation sequences follow W200 0.51 0.87 0.9951 W850 4.48 0.58 0.9994
the order S200 4 F200 4 W200 E Z200 E A200 4 AC 4 T200 Z200 0.47 0.87 0.9950 Z850 3.35 0.58 0.9982

This journal is © the Owner Societies 2014 Phys. Chem. Chem. Phys.

Paper
Published on 03 October 2014. Downloaded by UNIVERSIDAD DE EXTREMADURA on 21/10/2014 10:40:58.
Fig. 6 Electrical conductivity versus volume for nanocomposites of (a) series 1 and (b) series 2.
needed for reaching a given conductivity value is much broader
for composites prepared at 200 1C. Despite the foregoing, it is
also seen that the sensitivity of s to the volume change is as a
rule quite similar for AC and all the twelve nanocomposites,
given that their corresponding s–V curves are rather similarly
shaped and almost parallel. Furthermore, one observes in
Fig. 6(a) and (b) that the evolution of s with decreasing V
exhibits a good linear relationship for all the samples at P 4
100 kPa. Since for most hybrid materials the volume change is
greater at P values lower than the above, the results indicate
that only after the samples have undergone an initial volume
reduction is further compression more effective in increasing
their electrical conductivity. After such a reduction, small
changes in the sample volume as a result of increasing pressure
lead to a great enhancement in s.
Fig. 5 and 6 show that the relative magnitude of s as a rule
depends on whether the plot is depicted against P or V. In the
case of T200 and Z200, as an example, s is nearly identical over
a wide V range, but significantly higher for T200 over almost the
entire P range. For T850 and W850, s is markedly higher for the
latter over the entire P range, whereas the opposite applies over
a wide V range.
3.3.3. Variation of r with mechanical work. The plot of
s against the mechanical work (MW, henceforth) allows analyzing
together the opposite effects of P and V on s. MW has been
estimated for each sample using the following equation:110
X
i   s–MW curves in Fig. 7(a) and (b) is extremely different from
MW ¼ pj  A  hj1  hj
j¼1
where A is the surface area of the piston and pj denotes the
pressure required to cause a change of (hj1  hj) in the height
of the sample bed. Thus, the value of MW is identified with the
resistance of a given sample to be reduced in volume by
compression. The above equation is fully applicable in the case
of a homogeneous medium. Nevertheless, its applicability for
non-homogeneous media, such as AC and the hybrid materials
studied in the present work, is very questionable as the
Phys. Chem. Chem. Phys.
View Article Online
PCCP
significant effects of orientation of the grains, and of pressure
inhomogeneities in the sample bed, are obviated.41 As reported
by Celzard et al.,44 such effects depend on several factors,
especially the kind of carbonaceous material (e.g. activated
carbon, powder graphite or carbon black, among others), the
bed height and the applied pressure. In the present study, for
obtaining comparable results, only a narrow range of initial
heights of the sample column between 2.5 and 4.0 mm was used,
while the pressure range was the same for AC and all twelve
nanocomposites. The evolution of s with MW is illustrated in
Fig. 7(a) and (b).
As can be seen, for each sample, the s–MW curve is faintly
similar to the s–P plot (Fig. 5(a) and (b)) but it does not bear any
resemblance to the s–V plot (Fig. 6(a) and (b)), thus suggesting
that the applied pressure rather than the sample volume is the
predominant factor responsible for the variation of s with MW.
Such behaviour is related to the fact that, as previously shown
in Fig. 3(a) and (b), a large increase in P is usually required to
produce a relatively small reduction in the sample volume. As a
rule, for AC and both series of composites, s sharply increases
until a certain value of MW is reached and, then, the rise
flattens out. This occurs in the narrow range between 4  104
and 5  104 J for AC and most composites. In addition, for a
given metal oxide precursor, the increase in heat treatment
temperature from 200 to 850 1C does not significantly modify the
shape of the s–MW curves for the resulting hybrid materials.
However, it must be pointed out that the distribution of the
(7)
that obtained for the s–P plots in Fig. 5(a) and (b), where some
of the curves, especially those for composites of series 2, lie very
close to each other. These results clearly indicate that s is more
sensitive to the increase in MW than in P and, therefore, the
AC and the prepared hybrid materials are better discriminated
by using the extrinsic factor MW. Indeed, for samples S200
and W200, a conductivity of E70 S m1 is achieved with a P E
300 kPa, whereas MW values of 1.5  104 and 3.0  104 J,
respectively, must be applied to reach the same conductivity
value. In the same way, under a P E 756 kPa nanocomposites
This journal is © the Owner Societies 2014

PCCP
Published on 03 October 2014. Downloaded by UNIVERSIDAD DE EXTREMADURA on 21/10/2014 10:40:58.
Fig. 7 Electrical conductivity versus mechanical work for nanocomposites of (a) series 1 and (b) series 2.
A850 and Z850 exhibit a conductivity of E155 S m1, which
requires MW values of 7.6  104 and 1.0  103 J, respectively.
3.3.4. Variation of r with q. The variation of s with r for
composites of series 1 and 2 is depicted in Fig. 8(a) and (b),
respectively. As expected, for each hybrid material, the s–r plot
closely resembles the s–V plot. Unlike the behaviour observed
for other carbon materials, mainly carbon blacks,41,110 the total
variation in s with increasing r of the samples by compression
appears not to be dependent on the density of the nanocompo-
sites. In the case of samples prepared at 200 1C, the increase in
s is roughly as follows: AC 4 T200 4 S200 E W200 4 Z200 4
A200 44 F200, while for those prepared at 850 1C it varies in
the order: W850 4 AC 4 S850 E T850 4 A850 E Z850 4 F850.
In both cases, a clear correlation between the variation sequences
of s with r and r with P is not found. Furthermore, s increases
steeply and almost linearly with r at P values above ca. 100 kPa,
the slope of the fitted lines being quite comparable for AC and
all the hybrid materials. In this regard, an analogous linear
relationship between s and r has also been reported elsewhere46
for compacts of several carbon materials, including graphene,
multiwalled carbon nanotubes and graphite powder. This varia-
tion of s with increasing r of the samples may be explained as
follows. The electrical conductivity is strongly determined by
the number of effective electrical contacts that are established
between neighbouring particles and grains of the samples.
On this basis, one may conclude that for each sample the number
Fig. 8 Electrical conductivity versus density for nanocomposites of (a) series 1 and (b) series 2.
This journal is © the Owner Societies 2014
View Article Online
Paper
of electrical contacts, and thereby the conductivity, should be
primarily a function of its density, among other factors, since the
greater the density, the closer the particles and the less the void
space between them, which strongly contributes to the bulk
electrical resistance of the material. Under compression, the
neighbouring particles of the sample are forced to approach
each other, thus increasing both the density and the number of
electrical contacts. As shown in Fig. 8(a) and (b), a small increase
in the density of AC and the hybrid materials leads to a large
enhancement in electrical conduction, which clearly reveals the
loose packing of the sample particles just prior to starting the
compaction experiments.
As previously shown for other carbon materials,41,44,111 the
morphology of the samples may also play a major role in their
electrical conductivity. Nevertheless, the influence of this factor
on the conductivity of the prepared hybrid materials has been
omitted in the present work, since the particle size distribution
is almost identical for AC and all composites with an average
particle size of E1.5 mm.
In order to simplify and address the problem of the electrical
conduction in the prepared nanocomposites under moderate
compression, they have been modelled as particulate materials
consisting of AC grains partially coated by nanoparticles of
metal oxides, which may also be accompanied by hydroxides,
oxyhydroxides, metals in the elemental state and even carbides,
depending both on the metal oxide precursor and the heat
Phys. Chem. Chem. Phys.

Paper
treatment temperature (see Table 2). This simplified model is in
line with the mean crystallite size estimated for the various
supported metal phases (see Table 2 again), which ranges from
1.88 nm for Al2O3 in A200 to 20.80 nm for ZnO in Z200, and
between 1.86 nm for TiO2-rutile in T850 and 89.62 nm for Sn in
S850. A few previous studies dealing with the electrical behaviour
Published on 03 October 2014. Downloaded by UNIVERSIDAD DE EXTREMADURA on 21/10/2014 10:40:58.
of lightly compacted metal powders from both theoretical and
experimental standpoints56,112–115 have proved that the very thin
films coating the metallic powder particles do not affect the
mechanical properties, but substantially alter the electrical pro-
perties of the powder mass, particularly at the early stages of the
compression experiment. Such an assertion seems to be fully valid
for the prepared carbon-based hybrid materials since, as stated
above, their mechanical properties (i.e., densification and defor-
mation behaviours) are quite similar to those of the raw AC, while
their electrical conductivity is largely controlled by the different
nanoparticles supported on the AC surface. However, several
metal phases can be found together at different proportions in
a single nanocomposite (as an illustrative example, consider the
case of the sample F200), which usually makes a very difficult task
to accurately know the actual electrical conductivity of the nano-
particles covering the AC grains.
The electrical contacts will be more effective, and then will
allow electron transport easily, when they are established between
points of carbon surfaces belonging to neighbouring composite
particles, as one infers from the greatest conductivity values
measured for the pristine AC. The metal oxide, hydroxide and
oxyhydroxide nanoparticles in the prepared hybrid materials act as
insulating thin layers due to their much lower electrical conduc-
tivity values as compared to the carbon support. Therefore, the
contacts between close composite particles involving the aforesaid
insulating films will be less effective as paths for the electron
transport, thus increasing the contact resistance and decreasing
the bulk electrical conductivity of the hybrid material. Conversely,
highly conductive nanoparticles of metals in the elemental state
and metal carbides detected in composites F850, S850 and W850
are likely to favour the electrical current flow under compression,
as long as they are located in the small contact area between two
sample particles. In this connection, it should be noted that for the
pair of composites F850 and S850 elemental iron and tin are
accompanied by their corresponding oxides Fe2O3 and SnO2.
Of both opposite effects of metal oxides and elemental metals
on the electrical properties of the prepared hybrid materials, the
former appears to be stronger so that the bulk conductivity of the
above samples is still noticeably lower than for AC. In contrast, for
W850 only metallic tungsten and tungsten carbide are detected,
whose conducting properties (see Table 2) redound to the improve-
ment of the bulk electrical conductivity for this nanocomposite as
compared to the carbon support. In brief, the great variety of
supported nanoparticles may be considered to act as electrical
switches either hindering or favouring the free electron transport
between the AC cores of composite particles in contact under
compression. Which of the two opposite behaviours occurs is
ultimately determined by the intrinsic conductivity of the nano-
particle, which depends on its chemical nature and, thereby, on
the metal oxide precursor and the heat treatment temperature.
Phys. Chem. Chem. Phys.
View Article Online
PCCP
Electron tunnelling is by far the predominant electrical
transport mechanism in granular and powder carbon materials
under compression. However, other additional and specific
conduction mechanisms may also contribute to the bulk electrical
conductivity of the hybrid materials, depending on the chemical
nature of the supported nanoparticles. Among them, those con-
cerning the various iron oxides have been the most widely inves-
tigated. Thus, several models have been described to explain the
electrical conductivity in a-Fe2O3 under moderate compression; the
most accepted one assumes conduction to take place primarily in
the d levels of Fe3+ ions either with or without additional conduc-
tion in the sp-bands of oxide ions.116,117 Kaneko et al. proposed
a mechanism for the electrical transport in lightly compacted
a-FeO(OH) powders involving the hopping of an electron from
the t2g orbital of a Fe2+ ion to the t2g orbital of an adjacent Fe3+
ion, due to the slight overlapping of both orbitals.118
4. Conclusions
Two series of AC–metal oxide (Al2O3, Fe2O3, SnO2, TiO2, WO3
and ZnO) nanocomposites were prepared by wet impregnation,
oven-drying at 120 1C and subsequent heat treatment at 200 or
850 1C in an inert atmosphere. The electrical conductivity of AC
and the resulting hybrid materials was studied under moderate
compression. The mechanical properties of the nanocomposites
are essentially determined by the carbon matrix. Both the decrease
in volume and the increase in density under compression are only
important at pressure values lower than 100 kPa for all samples.
Unlike the mechanical properties, the bulk electrical conductivity
of the hybrid materials is strongly influenced by the intrinsic
conductivity, mean crystallite size, content and chemical nature
of the supported nanoparticles, which ultimately depend on the
metal oxide precursor and the heat treatment temperature.
In this regard, for each precursor, conductivity improves with
temperature rise due to the increase in the crystallite size and,
in the case of Fe2O3, SnO2 and WO3 samples, to the formation
of the corresponding metals in the elemental state and even
metal carbides. Depending on their chemical nature, supported
nanoparticles behave as electrical switches either hindering
or favouring the free electron transport between the AC cores
belonging to composite particles in contact under compression.
Conductivity values measured for the nanocomposites are as a
rule lower than those for AC, all of them falling in the range of
semiconductor materials. Of the opposite effects of pressure
and volume comprised in the variation of the conductivity with the
mechanical work, the effect of pressure appears to predominate,
as deduced from the slight resemblance between the plots of
conductivity versus pressure and mechanical work.
Acknowledgements
Financial support by Gobierno de Extremadura and European
FEDER Funds is gratefully acknowledged. A. Barroso-Bogeat
thanks Spanish Ministerio de Educación, Cultura y Deporte for
the concession of a FPU grant (AP2010-2574).
This journal is © the Owner Societies 2014

PCCP
References
1 F. Rodrı́guez-Reinoso, Carbon, 1998, 36, 159–175.
2 H. Jüntgen, Fuel, 1986, 65, 1436–1446.
3 Y. Yang, K. Chiang and N. Burke, Catal. Today, 2011, 178,
197–205.
Published on 03 October 2014. Downloaded by UNIVERSIDAD DE EXTREMADURA on 21/10/2014 10:40:58.
4 P. Serp and J. L. Figueiredo, Carbon Materials for Catalysis,
John Wiley & Sons, Hoboken, New Jersey, 2009.
5 M. Goyal, Nonenvironmental Industrial Applications of
Activated Carbon Adsorption, in Novel Carbon Adsorbents,
ed. J. M. D. Tascón, Elsevier, Amsterdam, 2012, pp. 605–638.
6 Q. Huang, X. Wang, J. Li, C. Dai, S. Gamboa and P. J. Sebastian,
J. Power Sources, 2007, 164, 425–429.
7 J. H. Park, O. O. Park, K. H. Shin, C. S. Jin and J. H. Kim,
Electrochem. Solid-State Lett., 2002, 5, H7–H10.
8 M. Seredych, O. Mabayoje, M. M. Koleśnik, V. Krstić and
T. J. Bandosz, J. Mater. Chem., 2012, 22, 7970–7978.
9 Q. L. Zhang, Y. C. Lin, X. Chen and N. Y. Gao, J. Hazard.
Mater., 2007, 148, 671–678.
10 C. S. Castro, M. C. Guerreiro, M. Gonçalves, L. C. A.
Oliveira and A. S. Anastácio, J. Hazard. Mater., 2009, 164,
609–614.
11 A. Guha, W. Lu, T. A. Zawodzinski Jr and D. A. Schiraldi,
Carbon, 2007, 45, 1506–1517.
12 M. Roosta, M. Ghaedi and M. Mohammadi, Powder Technol.,
2014, 267, 134–144.
13 S. Pal, Y. Kyung Tak, J. Joardar, W. Kim, J. Eun Lee, M. Soo
Han and J. Myong Song, J. Nanosci. Nanotechnol., 2009, 9,
2092–2103.
14 H. Tamai, U. Nobuaki and H. Yasuda, Mater. Chem. Phys.,
2009, 114, 10–13.
15 P. Gorria, M. Sevilla, J. A. Blanco and A. B. Fuertes, Carbon,
2006, 44, 1954–1957.
16 M. Qinghan, L. Ling, S. Huaihe and L. Licheng, J. Mater.
Sci., 2004, 39, 3149–3150.
17 A. Barroso-Bogeat, C. Fernández-González, M. Alexandre-
Franco and V. Gómez-Serrano, Activated carbon as a metal
oxide support: a review, in Activated Carbon: Classifications,
Properties and Applications, ed. J. F. Kwiatkowski, Nova Science
Publishers, New York, 2011, pp. 297–318.
18 R. Sanjinés, M. D. Abad, C. Vâju, R. Smajda, M. Mionić and
A. Magrez, Surf. Coat. Technol., 2011, 206, 727–733.
19 L. Wei and G. Yushin, Nano Energy, 2012, 1, 552–565.
20 M. Zhi, C. Xiang, J. Li, M. Li and N. Wu, Nanoscale, 2013, 5,
72–88.
21 Y. P. Wu, E. Rahm and R. Holze, J. Power Sources, 2003,
114, 228–236.
22 S. Goriparti, E. Miele, F. De Angelis, E. Di Fabrizio, R. P.
Zaccaria and C. Capligia, J. Power Sources, 2014, 257,
421–443.
23 E. Antolini, Appl. Catal., B, 2009, 88, 1–24.
24 V. V. Malyshev, Russ. J. Phys. Chem. A, 2009, 83, 1940–1945.
25 Y. S. Kim, S. C. Ha, K. Kim, H. Yang, S. Y. Choi and Y. T. Kim,
Appl. Phys. Lett., 2005, 86, 213105.
26 M. I. Baraton, L. Merhari, H. Ferkel and J. F. Castagnet,
Mater. Sci. Eng., 2002, 19, 315–321.
This journal is © the Owner Societies 2014
View Article Online
Paper
27 A. V. Chadwick, R. M. Geatches and J. D. Wright, Philos.
Mag. A, 1991, 64, 999–1010.
28 T. Ates, C. Tatar and F. Yakuphanoglu, Sens. Actuators, A,
2013, 190, 153–160.
29 M. I. Baraton and L. Merhari, Mater. Res. Soc. Symp. Proc.,
2004, 828, 191–196.
30 S. Mrozowski, Proceedings 3rd Biennal Carbon Conference,
Pergamon Press, Buffalo, 1957, p. 495.
31 R. Holm, Electric contacts, H. Geber, Stockholm, 1946.
32 R. Holm and E. Holm, Electric contacts: theory and applica-
tion, Springer-Verlag, Berlin, 4th edn, 1967.
33 K. J. Euler, R. Kirchhof and H. Metzendorf, Mater. Chem.,
1979, 4, 611–630.
34 C. Giraud, J. P. Clerc and E. Guyon, Powder Technol., 1983,
35, 107–111.
35 F. I. Zorin, Inorg. Mater., 1986, 22, 47–49.
36 N. Deprez and D. S. McLachlan, J. Phys. D: Appl. Phys.,
1988, 21, 101–107.
37 A. Gervois, M. Ammi, T. Travers, D. Bideau, J. C. Messager
and J. P. Troadec, Physica A, 1989, 157, 565–569.
38 J. P. Troadec and D. Bideau, Onde Electr., 1991, 71, 30–33.
39 K. Kinoshita, Carbon. Electrochemical and physicochemical
properties, Wiley, New York, 1988.
40 H. Braun and P. Herger, Mater. Chem., 1982, 7, 787–802.
41 J. Sánchez-González, A. Macı́as-Garcı́a, M. F. Alexandre-
Franco and V. Gómez-Serrano, Carbon, 2005, 43, 741–747.
42 F. Skaupy and O. Kantrowicz, Metallwirtsch., Metallwiss.,
Metalltech., 1931, 10, 45–47.
43 A. Espinola, P. Mourente Miguel, M. Roedel Salles and
A. Ribeiro Pinto, Carbon, 1986, 24, 337–341.
44 A. Celzard, J. F. Marêché, F. Payot and G. Furdin, Carbon,
2002, 40, 2801–2815.
45 K. J. Euler, J. Power Sources, 1978, 3, 117–136.
46 B. Marinho, M. Ghislandi, E. Tkalya, C. E. Koning and
G. De With, Powder Technol., 2012, 221, 351–358.
47 S. Marinkovic, C. Suznjevic and M. Djordjevic, Phys. Status
Solidi A, 1971, 4, 743–754.
48 J. A. Dodds, J. Colloid Interface Sci., 1980, 77, 317–327.
49 R. Ben Aı̈m and P. le Goff, Powder Technol., 1968, 2, 1–12.
50 J. P. Troadec, A. Gervois, D. Bideau and L. Oger, J. Phys. C:
Solid State Phys., 1987, 20, 993–1004.
51 A. Bertei and C. Nicolella, Powder Technol., 2011, 213, 100–108.
52 A. Barroso-Bogeat, M. Alexandre-Franco, C. Fernández-
González and V. Gómez-Serrano, Fuel Process. Technol.,
2014, 126, 95–103.
53 A. Barroso-Bogeat, M. Alexandre-Franco, C. Fernández-
González and V. Gómez-Serrano, J. Mater. Sci. Technol.,
submitted.
54 B. D. Cullity, Elements of X-ray diffraction, Addison-Wesley,
Reading, MA, 1959.
55 J. I. Langford and A. J. C. Wilson, J. Appl. Crystallogr., 1978,
11, 102–113.
56 J. M. Montes, F. G. Cuevas, J. Cintas and P. Urban, Appl.
Phys. A: Mater. Sci. Process., 2011, 105, 935–947.
57 N. Tsuda, K. Nasu, A. Fujimori and K. Siratori, Electronic
Conduction in Oxides, Springer, New York, 2nd edn, 2000.
Phys. Chem. Chem. Phys.

Paper
58 W. T. Li, D. R. McKenzie, W. D. McFall, Q. C. Zhang and
W. Wiszniewski, Solid-State Electron., 2000, 44, 1557–1562.
59 A. A. Akl, Appl. Surf. Sci., 2004, 221, 319–329.
60 N. Guskos, G. J. Papadopoulos, V. Likodimos, S. Patapis,
D. Yarmis, A. Przepiera, K. Przepiera and J. Majszczyk,
Mater. Res. Bull., 2002, 37, 1051–1061.
Published on 03 October 2014. Downloaded by UNIVERSIDAD DE EXTREMADURA on 21/10/2014 10:40:58.
61 D. Reisinger, P. Majewski, M. Opel, L. Alff and R. Gross,
Appl. Phys. Lett., 2004, 85, 4980.
62 A. V. Moholkar, S. M. Pawar, K. Y. Rajpure, C. H. Bhosale
and J. H. Kim, Appl. Surf. Sci., 2009, 255, 9358–9364.
63 K. Krishnakumar, N. Pinna, K. Perumal and R. Jayaprakash,
AIP Conf. Proc., 2008, 1004, 122–125.
64 K. W. Vogt, P. A. Kohl, W. B. Carter, R. A. Bell and L. A.
Bottomley, Surf. Sci., 1994, 301, 203–213.
65 M. G. Hutchins, O. Abu-Alkhair, M. M. El-Nahass and
K. Abdel-Hady, J. Non-Cryst. Solids, 2007, 353, 4137–4142.
66 M. M. El-Nahass, H. A. M. Ali, M. Saadeldin and M. Zaghllol,
Physica B, 2012, 407, 4453–4457.
67 L. Beluze, J. C. Badot, R. Weil and V. Lucas, J. Phys.
Chem. B, 2006, 110, 7304–7308.
68 H. Wang, C. Li, H. Zhao, R. Li and J. Liu, Powder Technol.,
2013, 239, 266–271.
69 F. G. de Souza Jr, J. A. Marins, J. C. Pinto, G. E. de Oliveira,
C. M. Rodrigues and L. M. T. R. Lima, J. Mater. Sci., 2010,
45, 5012–5021.
70 E. A. Brandes and G. B. Brook, Smithells Metals Reference
Book, Butterworth Heinemann, Oxford, 7th edn, 1992.
71 T. Dittrich, J. Weidmann, F. Koch, I. Uhlendorf and
I. Lauermann, Appl. Phys. Lett., 1999, 75, 3980–3982.
72 H. O. Pierson, Handbook of Chemical Vapour Deposition
(CVD): Principles, Technology, and Applications, William
Andrew Inc., New York, 2nd edn, 1999.
73 S. Torquato and M. D. Rintoul, Phys. Rev. Lett., 1995, 75,
4067–4070.
74 S. J. Gregg and K. S. W. Sing, Adsorption, surface area and
porosity, Academic Press, London, 2nd edn, 1982.
75 S. Brunauer, P. H. Emmet and E. Teller, J. Am. Chem. Soc.,
1938, 60, 309–319.
76 M. M. Dubinin, in Progress in Surface and Membrane
Science, ed. J. F. Danielli, M. D. Rosenberg and D. A.
Cadenhead, Academic Press, New York, 1975, vol. 9, pp. 1–70.
77 D. Pantea, H. Darmstadt, S. Kaliaguine, L. Sümmchen and
C. Roy, Carbon, 2001, 39, 1147–1158.
78 M. Smisek and S. Cerny, Active Carbon: Manufacture,
Properties and Applications, Elsevier, Amsterdam, 1979.
79 C. A. León y León, Adv. Colloid Interface Sci., 1998, 76–77,
341–372.
80 K. K. Nanda, Phys. Lett. A, 2012, 376, 3301–3302.
81 W. D. Callister Jr, Introducción a la Ciencia e Ingenierı́a de
los Materiales, Reverté, Barcelona, 1996, vol. 2.
82 L. J. Kennedy, J. Judith Vijaya and G. Sekaran, Mater. Chem.
Phys., 2005, 91, 471–476.
83 T. Adinaveen, L. J. Kennedy, J. Judith Vijaya and G. Sekaran,
J. Ind. Eng. Chem., 2013, 19, 1470–1476.
84 W. M. Daud, M. Badri and H. Mansor, J. Appl. Phys., 1990,
67, 1915–1917.
Phys. Chem. Chem. Phys.
View Article Online
PCCP
85 A. Barroso-Bogeat, M. Alexandre-Franco, C. Fernández-
González, A. Macı́as-Garcı́a and V. Gómez-Serrano,
Microporous Mesoporous Mater., 2014, DOI: 10.1016/
j.micromeso.2014.07.023.
86 M. Koós, S. H. S. Moustafa, E. Szilágyi and I. Pócsik,
Diamond Relat. Mater., 1999, 8, 1919–1926.
87 S. H. Moustafa, M. Koós and I. Pócsik, J. Non-Cryst. Solids,
1998, 227–230, 1087–1091.
88 Y. Hishiyama, Y. Kaburagi and A. Ono, Carbon, 1979, 17,
265–276.
89 Y. Inoue, M. Sugano, S. Kawakami, M. Kitano,
K. Nakajima, H. Kato and M. Hara, Bull. Chem. Soc. Jpn.,
2013, 86, 45–50.
90 J. B. Donnet, Carbon, 1994, 32, 1305–1310.
91 G. D. Zhan and A. K. Mukherjee, Int. J. Appl. Ceram.
Technol., 2004, 1, 161–171.
92 Y. Sakata, A. Muto, M. A. Uddin, M. Tanihara, K. Harino,
J. Takada and Y. Kusano, J. Jpn. Soc. Powder Powder Metall.,
1998, 45, 806–811.
93 Y. Zhu, L. Zhang, F. M. Schappacher, R. Pöttgen,
J. Shi and S. Kaskel, J. Phys. Chem. C, 2008, 112,
8623–8628.
94 Y. Liu and H. Yan, Mater. Lett., 2011, 65, 1063–1065.
95 Y. Ao, J. Xu, X. Shen, D. Fu and C. Yuan, J. Hazard. Mater.,
2008, 160, 295–300.
96 C. S. Castro, M. C. Guerreiro, M. Gonçalves, L. C. A.
Oliveira and A. S. Anastácio, J. Hazard. Mater., 2009, 164,
609–614.
97 Q. L. Zhang, Y. C. Lin, X. Chen and N. Y. Gao, J. Hazard.
Mater., 2007, 148, 671–678.
98 A. Quintanilla, N. Menéndez, J. Tornero, J. A. Casas and
J. J. Rodrı́guez, Appl. Catal., B, 2008, 81, 105–114.
99 K. L. López-Maldonado, P. de la Presa, M. A. de la Rubia,
P. Crespo, J. de Frutos, A. Hernando, J. A. Matutes Aquino
and J. T. Elizalde Galindo, J. Nanopart. Res., 2014, 16, 2482.
100 S. L. Patil, M. A. Chougule, S. G. Pawar, B. T. Raut, S. Sen
and V. B. Patil, J. Alloys Compd., 2011, 509, 10055–10061.
101 M. J. Iqbal, N. Yaqub, B. Sepiol and B. Ismail, Mater. Res.
Bull., 2011, 46, 1837–1842.
102 N. Ponpandian and A. Narayanasamy, J. Appl. Phys., 2002,
92, 2770–2778.
103 N. Tigau, V. Ciupina, G. Prodan, G. I. Rusu, C. Gheorghies
and E. Vasile, J. Optoelectron. Adv. Mater., 2003, 5, 907–912.
104 R. S. Rao, A. B. Walters and M. A. Vannice, J. Phys. Chem. B,
2005, 109, 2086–2092.
105 R. N. Viswanath and S. Ramasamy, Colloids Surf., A, 1998,
133, 49–56.
106 A. W. Czanderna, C. N. Ramachandra Rao and J. M. Honig,
Trans. Faraday Soc., 1958, 54, 1069–1073.
107 B. Tryba, A. W. Morawski and M. Inagaki, Appl. Catal., B,
2003, 41, 427–433.
108 Y. Li, S. Zhang, Q. Yu and W. Yin, Appl. Surf. Sci., 2007, 253,
9254–9258.
109 Y. X. Chen, B. H. Luo, B. Y. Zhao, Y. J. Lai, H. Z. Wang,
M. Y. Gao, W. X. Zhang and K. B. Chen, J. Inorg. Organomet.
Polym., 2012, 22, 90–96.
This journal is © the Owner Societies 2014

PCCP
110 N. Probst and E. Grivei, Carbon, 2002, 40, 201–205.
111 D. Pantea, H. Darmstadt, S. Kaliaguine, S. Blacher and
C. Roy, Rubber Chem. Technol., 2002, 75, 691–700.
112 J. M. Montes, F. G. Cuevas and J. Cintas, Metall. Mater.
Trans. B, 2007, 38, 957–964.
113 J. M. Montes, F. G. Cuevas and J. Cintas, Granular Matter,
Published on 03 October 2014. Downloaded by UNIVERSIDAD DE EXTREMADURA on 21/10/2014 10:40:58.
2011, 13, 439–446.
This journal is © the Owner Societies 2014
View Article Online
Paper
114 T. J. Garino, J. Mater. Res., 2002, 17, 2691–2697.
115 P.-G. De Gennes, Europhys. Lett., 1996, 35, 145–149.
116 F. J. Morin, Phys. Rev., 1954, 93, 1195–1199.
117 E. Knittle and R. Jeanloz, Solid State Commun., 1986, 58,
129–131.
118 K. Kaneko and K. Inouye, Bull. Chem. Soc. Jpn., 1974, 47,
1139–1142.
Phys. Chem. Chem. Phys.